Would be nice to see implicit permissions in @claudeai , ie if I say "fetch blog post https://linktopost and then do x" the next turn should not be Claude asking for permission to fetch the blog post.
Yup, platform activity is surging. There were 1 billion commits in 2025. Now, it's 275 million per week, on pace for 14 billion this year if growth remains linear (spoiler: it won't.)
GitHub Actions has grown from 500M minutes/week in 2023 to 1B minutes/week in 2025, and now 2.1B minutes so far this week.
So we're pushing incredibly hard on more CPUs, scaling services, and strengthening GitHubโs core features.
And as a fine purveyor of hand-crafted shit code for many years, I'm not gonna weigh in on that. ๐คฃ
Stripe Projects(https://t.co/0RqUqOTUcp) is interesting but this idea and moment seems ripe and more useful as a protocol that uses unrestricted money, BTC.
Starting to think that when people say "claude is getting worse" they may have just hit that level of complexity in their project where claude is starting to panic and make random changes just like any other junior dev thrown in the deep end.
@alexhillman Yea this is a simplification (my prompt has the details of my specific scenario). Just confirming there are different use cases and the execution model for skills vs mcp is different.
@alexhillman I asked it which was a better choice for when I want the agent to deterministically perform an action (make an api call) in response to a prompt
I also think we should change โDaylight Savings Timeโ back to its original name: War Time.
Letโs be honest with ourselves about the functions our government does and then it will be clearer when we can move on.